supraj.dev
Kubernetes

Kubernetes requests and limits

How does Kubernetes decide where my app runs, and what happens when it uses too much?

28 s loop7 sentencesSilent, read alongScene: Journey
Read the story, sentence by sentence 7 sentences

What you'll be able to say

  1. The kube-scheduler only places a pod on a node whose allocatable capacity still covers the pod’s resource requests, regardless of actual usage.
  2. A container that goes over its CPU limit is throttled, not killed, so it keeps running but more slowly.
  3. Over its memory limit, a container may be killed by the kernel (OOMKilled) and then restarted according to the pod’s restartPolicy.